Hi. I have a dataset consisting of two groups. for Group A, n=40 and for Group B, n=134.
I've been getting insignificant results with them, but it doesn't seem likely according to the literature. I use t-tests, mann-whitney U, and correlation analysis for several variables. I will also do a multiple regression later.
My question is: Does the unequal size affect my results? and could it be why I'm getting these weird results from my data?
Your larger group has a smaller SD and your small group a larger one. You can test the difference with independent t-test. If the result is n.s. then it is n.s. Your data are yours, not from the literature.
Dear Maryam
1. Your question is not clear, and what is your problem?
2. Several statistical measures are required equal sample sizes, and some other are not required equal sample sizes.
3.How many random variables do you have, how many dependent variables and independent variables?
4. What is the relation between the random variables, i.e. related or not.
5. Regression analysis also required a linear relationship between the independent random variables and the dependent random variable.
6. Linear regression model/analysis is a technique used to predict the value of one quantitative variable by using its relationship with one or more additional quantitative variables.
7. Normality assumption is required by all parametric tests, where non-parametric are developed for problems not followed normal distribution.
8. Selecting any kind of statistical measures or any analysis technique should be based on your hypotheses and research questions and the above requirements.
8. Thus, you need to determine the research questions/hypotheses of your research.
Regards,
Zuhair
with sample size so different, you must calculate and compare VARIANCES. If variances are DIFFERENT then simple pooling is not advisable. Use weighted t-test [this is assuming your samples are normally distributed; do a normality test]. with such large samples I would fit relevant distribution and calculate Confidence Intervals. Outliers can affect Means badly ...
Dear Maryam
In addition to my previous answer, please refer to:
Snedecor, W.G. and Cochran,W.G. (1989). Statistical Methods, Blackwell.
Landau, S. and Everitt, B. S.(2004). A Handbook of Statistical analyses using SPSS.
Howitt, D. and Cramer, D.(2008). Introduction to SPSS.
Regards,
Zuhair
In such situations I like to see the populations and for this I use "Mountain or Folded quantile plot"
"---------------Folded percentile or Mountain plot------------ Reference: Krouwer JS and Monti KL.A simple, graphical method to evaluate laboratory assays. Eur J Clin Chem Clin Biochem. 1995 Aug;33(8):525-7."
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression